Data Mode

Premium

• Last updated Sep 4, 2026
• View as Markdown

The mode parameter allows you to control how an API request is fulfilled, including data freshness guarantees and credit usage. The mode parameter determines the response behavior used to fulfill each request.

Our API supports four freshness modes — live, delayed, eod, and historical — plus cached, which serves recently stored data at a reduced credit cost. These options balance immediacy, availability, and cost efficiency. Below is a detailed overview of each mode, including examples and recommended use-cases.

Premium Parameter

This parameter is available only on paid plans. Free and trial plans cannot change the data mode. They receive historical data — the last fully-closed session — and a request for any fresher mode (mode=live, mode=delayed, or mode=eod) returns 402 Payment Required. See Data Freshness for what each plan receives.

Live Mode

The live mode returns real-time data, providing the most current market information available. This mode is intended for time-sensitive use cases that require up-to-the-second pricing.

note

mode=live returns the freshest data your entitlements allow. It does not shorten an exchange-imposed delay. See Data Freshness for the freshness of each endpoint.

Pricing for Live Mode

• Quotes: 1 credit per symbol included in the response that has quote data (bid/ask/mid/last).
• Candles: 1 credit per 1,000 candles included in the response.
• Bulk Candles: 1 credit per symbol* included in the response.
• Other Endpoints: 1 credit per response.

Requesting Live Data

To request real-time data, append mode=live to your API call or omit the parameter entirely. If the mode parameter is not specified, live mode is used by default for paid accounts.

GET https://api.marketdata.app/v1/options/chain/SPY/?mode=live
GET https://api.marketdata.app/v1/options/chain/SPY/

Both requests are valid and return the most recent data available for the specified symbol.

Cached Mode

The cached mode returns recently stored data that may be a few seconds to several minutes old. This mode is designed to reduce credit usage when retrieving large amounts of quote data.

Data freshness is not guaranteed. Symbols with higher usage across Market Data customers tend to be refreshed more frequently.

Pricing for Cached Mode

• Quotes: 1 credit per request, regardless of the number of symbols.
• Historical Quotes: Unavailable
• Candles: Unavailable
• Bulk Candles: Unavailable
• Other Endpoints: Unavailable

This makes cached mode particularly cost-effective for bulk quote retrieval using endpoints such as Option Chain and Bulk Stock Quotes.

Requesting Cached Data

GET https://api.marketdata.app/v1/options/chain/SPY/?mode=cached

This request attempts to fulfill the response from Market Data’s cache. If cached data is available, the response is returned at a reduced credit cost.

Using maxage with Cached Mode

The maxage parameter can be combined with mode=cached to control how old cached data may be before it is considered unusable.

How maxage Works

• Purpose: Sets a maximum acceptable age for cached data
• Behavior: If no cached data exists within the specified age window, the API returns an empty response with no credit charge
• Format: Accepts absolute datetimes or relative durations (e.g. 1h, 5min)

Example Scenarios

Assume the most recent cached options chain for AAPL is 3 minutes old.

Scenario 1: Data within maxage limit

GET https://api.marketdata.app/v1/options/chain/AAPL/?mode=cached&maxage=5min

• Result: 203 response with data (1 credit charged)
• Reason: Cached data is newer than 5 minutes

Scenario 2: Data exceeds maxage limit

GET https://api.marketdata.app/v1/options/chain/AAPL/?mode=cached&maxage=1min

• Result: 204 empty response (0 credits charged)
• Reason: Cached data exceeds the specified age limit

Strategic Usage

Using maxage enables efficient fallback logic:

1. First attempt to retrieve cached data with a defined freshness threshold
2. If a 204 response is returned, selectively request live data for required symbols
3. This maximizes cost efficiency while preserving data freshness where it matters

Cached Mode Response Codes

When mode=cached is used, successful responses do not return 200 OK. Instead:

• 203 NON-AUTHORITATIVE INFORMATION – Request succeeded and was fulfilled from cache
• 204 NO CONTENT – No cached data available within constraints; no credits charged

204 is the deterministic cache-miss signal — it can only be returned by mode=cached (or mode=cache / mode=stale). For the full mapping of status codes across all modes, see Status Codes below.

Delayed Mode

The delayed mode returns data delayed by at least 15 minutes. It is available on paid plans (Starter and up), which may request delayed data explicitly.

For when delayed data crosses into "historical" (a fully-closed prior session) — and why that happens at different times for stocks (4:15 PM ET) vs options (9:30 AM ET the next trading day) — see Data Freshness.

Pricing for Delayed Mode

• Pricing is identical to live mode.

Requesting Delayed Data

GET https://api.marketdata.app/v1/options/chain/SPY/?mode=delayed

This request returns market data that is delayed by a minimum of 15 minutes.

EOD and Historical Modes

Two further modes return a closed session rather than a delay on the current one.

• mode=eod returns the most recent session whose 4:15 PM ET cutoff has passed. Requires a paid plan.
• mode=historical returns the most recent fully-closed session — the API waits for the next session to open before it counts a session as closed. This is what free and trial accounts receive, and the only mode they may request explicitly.

Pricing for EOD and Historical Modes

• Pricing is identical to live mode.

Requesting EOD or Historical Data

GET https://api.marketdata.app/v1/options/chain/SPY/?mode=eod
GET https://api.marketdata.app/v1/options/chain/SPY/?mode=historical

Mode Comparison

Feature	Live Mode	Cached Mode	Delayed Mode	EOD Mode	Historical Mode
Data Timeliness	Real-time	Seconds to minutes old	≥ 15 minutes delayed	Last session past 4:15 PM ET	Last fully-closed session
Pricing	1 credit per symbol with quote data	1 credit per request	Same as live	Same as live	Same as live
Ideal Use-Case	Time-sensitive decisions	Bulk quote retrieval	Non-time-sensitive applications	End-of-day reporting	Backtesting and research
Default Behavior	Paid accounts (if mode is omitted)	Must specify mode=cached	Must specify mode=delayed	Must specify mode=eod	Free & trial accounts
Paid Accounts Access	✅	✅	✅	✅	✅
Free/Trial Accounts Access	❌	❌	❌	❌	✅

Free and trial accounts receive historical data on every request. See Data Freshness for the freshness of each endpoint on each plan.

• Choose mode=live when immediate data freshness is required.
• Use mode=cached to reduce credit usage when working with large symbol sets.
• Select mode=delayed for applications where timing precision is not critical.
• Select mode=eod or mode=historical when a closed session is what you want.

Status Codes

Market Data uses HTTP status codes to communicate where a successful response came from. All three of 200, 203, and 204 are success responses — your client must accept all of them.

Status	Meaning	When it occurs
200 OK	Response was freshly fetched from the upstream provider.	Any mode, when no cache layer can satisfy the request.
203 Non-Authoritative Information	Response was served from a cache layer (Redis, database quote cache, response log, option-chain cache, etc.).	Any mode. Common during market hours regardless of mode=live, mode=delayed, or no mode specified. The body is identical in shape to a 200.
204 No Content	No cached data is available within the requested constraints. No credits charged.	Only when mode=cached (also mode=cache / mode=stale). Never returned by other modes.

Mode does not deterministically map to status code

A common (incorrect) assumption is that mode=live always returns 200 and mode=delayed always returns 203. Both can return either 200 or 203 depending on whether a cache layer can satisfy the request at the moment. Only mode=cached is deterministic — it returns 203 on cache hit or 204 on cache miss, never 200.

Handling 204 (cache miss on mode=cached)

When mode=cached returns 204, the typical pattern is to fall back to a live request. A short Python example:

import requests

def get_option_chain(token):
    url = "https://api.marketdata.app/v1/options/chain/AAPL/"
    headers = {"Authorization": f"Bearer {token}"}

    r = requests.get(url, params={"mode": "cached"}, headers=headers)
    if r.status_code in (200, 203):
        return r.json()
    if r.status_code == 204:
        # Cache miss — fall back to live (incurs normal live-mode credit cost)
        r = requests.get(url, params={"mode": "live"}, headers=headers)
        if r.status_code in (200, 203):
            return r.json()
    r.raise_for_status()

A few notes on the retry pattern:

• A retry with mode=live is billed at the full live-mode credit cost — see Pricing for Live Mode. The original 204 response is free.
• Cap the retry at one attempt. A subsequent 204 should not occur on mode=live, but a circuit breaker is still wise.
• If your plan does not include mode=cached access (Free/Trial), mode=cached requests return 402 Payment Required rather than 204. See 402: Payment Required.

For the full list of HTTP status codes returned by the API (including 4xx and 5xx), see Troubleshooting.